Layman’s Abstract for Applied Stochastic Models in Business and Industry article: An enhanced two-quantile Wilks methodology for engineering uncertainty analysis

Each week, we publish layman’s abstracts of new articles from our prestigious portfolio of journals in statistics. The aim is to highlight the latest research to a broader audience in an accessible format.
The article featured today is from Applied Stochastic Models in Business and Industry with the full article now available to read here.
van Dorp, JRShittu, ERabiti, CAn enhanced two-quantile Wilks methodology for engineering uncertainty analysisAppl Stochastic Models Bus Ind20221– 26. doi:10.1002/asmb.2729
In general, statistical estimation procedures regarding the uncertainty in a metric of interest are based on the form of its density function being a member of a specified family of distributions. Complex computational analyses involving an output metric of an engineering model with uncertain input parameters, however, result in that parametric form of the output metric’s uncertainty distribution being unknown. With the enhancements in computational facilities over the past decades, this has become more prevalent in engineering systems safety and design analyses, including nuclear and chemical engineering. Though assumptions are sometimes made on the nature of the distribution of such an output metric, it is quite important to note that deviations or discrepancies in assigning or specifying erroneous densities portend significant consequences. The solemn impacts are more pronounced especially in nuclear and chemical engineering where actual safety limits and regulatory acceptance criteria must be aligned. For example, estimating the “the right” tolerance is crucial in meeting the licensing requirements set forth by the Nuclear Regulatory Commission (NRC) in their specification of the limits of safety criteria. It is against this backdrop that this paper examines non-parametric statistical inference methods that do not require knowledge of the underlying density of the output parameter of interest. Central to this paper
is an enhancement of the tolerance interval estimation methods proposed in the seminal paper by S.S. Wilks in 1941 entitled “Determination of sample sizes for setting tolerance limits.” A tolerance interval is an interval within which a specified proportion of the output metric’s values in question falls, while exceeding a specified confidence level associated with that interval. The proposed methodology results in smaller sample sizes being evaluated compared to the procedures proposed by Wilks therein. Smaller sample sizes are desirable for computational efficiency and the reduction in cost associated with evaluations requiring the computational power offered by super computers.
More Details